Export qos_class sysfs attributes for the CXL memory device. The QoS clas
should show up as /sys/bus/cxl/devices/memX/ram/qos_class for the volatile
partition and /sys/bus/cxl/devices/memX/pmem/qos_class for the persistent
partition. The QTG ID is retrieved via _DSM after supplying the
calculated bandwidth and latency for the entire CXL path from device to
the CPU. This ID is used to match up to the root decoder QoS class to
determine which CFMWS the memory range of a hotplugged CXL mem device
should be assigned under.
While there may be multiple DSMAS exported by the device CDAT, the driver
will only expose the first QTG ID per partition in sysfs for now. In the
future when multiple QTG IDs are necessary, they can be exposed. [1]
[1]: https://lore.kernel.org/linux-cxl/167571650007.587790.10040913293130712882.stgit@djiang5-mobl3.local/T/#md2a47b1ead3e1ba08f50eab29a4af1aed1d215ab
Suggested-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/170319625698.2212653.17544381274847420961.stgit@djiang5-mobl3
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Restricted CXL host (RCH) downstream port AER information is not currently
logged while in the error state. One problem preventing the error logging
is the AER and RAS registers are not accessible. The CXL driver requires
changes to find RCH downstream port AER and RAS registers for purpose of
error logging.
RCH downstream ports are not enumerated during a PCI bus scan and are
instead discovered using system firmware, ACPI in this case.[1] The
downstream port is implemented as a Root Complex Register Block (RCRB).
The RCRB is a 4k memory block containing PCIe registers based on the PCIe
root port.[2] The RCRB includes AER extended capability registers used for
reporting errors. Note, the RCH's AER Capability is located in the RCRB
memory space instead of PCI configuration space, thus its register access
is different. Existing kernel PCIe AER functions can not be used to manage
the downstream port AER capabilities and RAS registers because the port was
not enumerated during PCI scan and the registers are not PCI config
accessible.
Discover RCH downstream port AER extended capability registers. Use MMIO
accesses to search for extended AER capability in RCRB register space.
[1] CXL 3.0 Spec, 9.11.2 - System Firmware View of CXL 1.1 Hierarchy
[2] CXL 3.0 Spec, 8.2.1.1 - RCH Downstream Port RCRB
Co-developed-by: Robert Richter <rrichter@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-12-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Now, that the Component Register mappings are stored, use them to
enable and map the HDM decoder capabilities. The Component Registers
do not need to be probed again for this, remove probing code.
The HDM capability applies to Endpoints, USPs and VH Host Bridges. The
Endpoint's component register mappings are located in the cxlds and
else in the port's structure. Duplicate the cxlds->reg_map in
port->reg_map for endpoint ports.
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
[rework to drop cxl_port_get_comp_map()]
Link: https://lore.kernel.org/r/20231018171713.1883517-8-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Pick up the first half of the RCH error handling series. The back half
needs some fixups for test regressions. Small conflicts with the PMU
work around register enumeration and setup helpers.
'struct cxl_dev_state' makes too many assumptions about the capabilities
of a CXL device. In particular it assumes a CXL device has a mailbox and
all of the infrastructure and state that comes along with that.
In preparation for supporting accelerator / Type-2 devices that may not
have a mailbox and in general maintain a minimal core context structure,
make mailbox functionality a super-set of 'struct cxl_dev_state' with
'struct cxl_memdev_state'.
With this reorganization it allows for CXL devices that support HDM
decoder mapping, but not other general-expander / Type-3 capabilities,
to only enable that subset without the rest of the mailbox
infrastructure coming along for the ride.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/168679260240.3436160.15520641540463704524.stgit@dwillia2-xfh.jf.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
CXL RAS capabilities must be enabled and accessible as soon as the CXL
endpoint is detected in the PCI hierarchy and bound to the cxl_pci
driver. This needs to be independent of other modules such as cxl_port
or cxl_mem.
CXL RAS capabilities reside in the Component Registers. For an RCH
this is determined by probing RCRB which is implemented very late once
the CXL Memory Device is created.
Change this by moving the RCRB probe to the cxl_pci driver. Do this by
using a new introduced function cxl_pci_find_port() similar to
cxl_mem_find_port() to determine the involved dport by the endpoint's
PCI handle. Plug this into the existing cxl_pci_setup_regs() function
to setup Component Registers. Probe the RCRB in case the Component
Registers cannot be located through the CXL Register Locator
capability.
This unifies code and early sets up the Component Registers at the
same time for both, VH and RCH mode. Only the cxl_pci driver is
involved for this. This allows an early mapping of the CXL RAS
capability registers.
Signed-off-by: Robert Richter <rrichter@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/20230622205523.85375-14-terry.bowman@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In order to move the RCH dport component register setup to cxl_pci the
base address must be stored in CXL device state (cxlds) for both
modes, RCH and VH. Store it in cxlds->component_reg_phys and use it
for endpoint creation.
Signed-off-by: Robert Richter <rrichter@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/20230622205523.85375-13-terry.bowman@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
For symmetry with the recent rename of ->dport_dev for a 'struct
cxl_dport', add the "_dev" suffix to the ->uport property of a 'struct
cxl_port'. These devices represent the downstream-port-device and
upstream-port-device respectively in the CXL/PCIe topology.
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/20230622205523.85375-6-terry.bowman@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The RCRB is extracted already during ACPI CEDT table parsing while the
data of this is needed not earlier than dport creation. This
implementation comes with drawbacks: During ACPI table scan there is
already MMIO access including mapping and unmapping, but only ACPI
data should be collected here. The collected data must be transferred
through a couple of interfaces until it is finally consumed when
creating the dport. This causes complex data structures and function
interfaces. Additionally, RCRB parsing will be extended to also
extract AER data, it would be much easier do this at a later point
during port and dport creation when the data structures are available
to hold that data.
To simplify all that, probe the RCRB at a later point during RCH
downstream port creation. Change ACPI table parser to only extract the
base address of either the component registers or the RCRB. Parse and
extract the RCRB in devm_cxl_add_rch_dport().
This is in preparation to centralize all RCRB scanning.
Signed-off-by: Robert Richter <rrichter@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/20230622205523.85375-2-terry.bowman@amd.com
Co-developed-by: Dan Williams <dan.j.williams@intel.com>
Link: https://lore.kernel.org/r/20230622205523.85375-3-terry.bowman@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Move cxl_await_media_ready() to cxl_pci probe before driver starts issuing
IDENTIFY and retrieving memory device information to ensure that the
device is ready to provide the information. Allow cxl_pci_probe() to succeed
even if media is not ready. Cache the media failure in cxlds and don't ask
the device for any media information.
The rationale for proceeding in the !media_ready case is to allow for
mailbox operations to interrogate and/or remediate the device. After
media is repaired then rebinding the cxl_pci driver is expected to
restart the capacity scan.
Suggested-by: Dan Williams <dan.j.williams@intel.com>
Fixes: b39cb1052a ("cxl/mem: Register CXL memX devices")
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/168445310026.3251520.8124296540679268206.stgit@djiang5-mobl3
[djbw: fixup cxl_test]
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Inject and Clear Poison commands are optionally supported by CXL
memdev devices and are intended for use in debug environments only.
Add debugfs attributes for user access.
Documentation/ABI/testing/debugfs-cxl describes the usage.
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/0c9ea8e671b8e58465d18722788b60d325c675c7.1681874357.git.alison.schofield@intel.com
Tested-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
When a boolean 'true' is written to this attribute the memdev driver
retrieves the poison list from the device. The list consists of
addresses that are poisoned, or would result in poison if accessed,
and the source of the poison. This attribute is only visible for
devices supporting the capability. The retrieved errors are logged
as kernel events when cxl_poison event tracing is enabled.
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/1081cfdc8a349dc754779642d584707e56db26ba.1681838291.git.alison.schofield@intel.com
Tested-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Unlike a CXL memory expander in a VH topology that has at least one
intervening 'struct cxl_port' instance between itself and the CXL root
device, an RCD attaches one-level higher. For example:
VH
┌──────────┐
│ ACPI0017 │
│ root0 │
└─────┬────┘
│
┌─────┴────┐
│ dport0 │
┌─────┤ ACPI0016 ├─────┐
│ │ port1 │ │
│ └────┬─────┘ │
│ │ │
┌──┴───┐ ┌──┴───┐ ┌───┴──┐
│dport0│ │dport1│ │dport2│
│ RP0 │ │ RP1 │ │ RP2 │
└──────┘ └──┬───┘ └──────┘
│
┌───┴─────┐
│endpoint0│
│ port2 │
└─────────┘
...vs:
RCH
┌──────────┐
│ ACPI0017 │
│ root0 │
└────┬─────┘
│
┌───┴────┐
│ dport0 │
│ACPI0016│
└───┬────┘
│
┌────┴─────┐
│endpoint0 │
│ port1 │
└──────────┘
So arrange for endpoint port in the RCH/RCD case to appear directly
connected to the host-bridge in its singular role as a dport. Compare
that to the VH case where the host-bridge serves a dual role as a
'cxl_dport' for the CXL root device *and* a 'cxl_port' upstream port for
the Root Ports in the Root Complex that are modeled as 'cxl_dport'
instances in the CXL topology.
Another deviation from the VH case is that RCDs may need to look up
their component registers from the Root Complex Register Block (RCRB).
That platform firmware specified RCRB area is cached by the cxl_acpi
driver and conveyed via the host-bridge dport to the cxl_mem driver to
perform the cxl_rcrb_to_component() lookup for the endpoint port
(See 9.11.8 CXL Devices Attached to an RCH for the lookup of the
upstream port component registers).
Tested-by: Robert Richter <rrichter@amd.com>
Link: https://lore.kernel.org/r/166993045621.1882361.1730100141527044744.stgit@dwillia2-xfh.jf.intel.com
Reviewed-by: Robert Richter <rrichter@amd.com>
Reviewed-by: Jonathan Camerom <Jonathan.Cameron@huawei.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
tl;dr: Clean up an unnecessary export and enable cxl_test.
An RCD (Restricted CXL Device), in contrast to a typical CXL device in
a VH topology, obtains its component registers from the bottom half of
the associated CXL host bridge RCRB (Root Complex Register Block). In
turn this means that cxl_rcrb_to_component() needs to be called from
devm_cxl_add_endpoint().
Presently devm_cxl_add_endpoint() is part of the CXL core, but the only
user is the CXL mem module. Move it from cxl_core to cxl_mem to not only
get rid of an unnecessary export, but to also enable its call out to
cxl_rcrb_to_component(), in a subsequent patch, to be mocked by
cxl_test. Recall that cxl_test can only mock exported symbols, and since
cxl_rcrb_to_component() is itself inside the core, all callers must be
outside of cxl_core to allow cxl_test to mock it.
Reviewed-by: Robert Richter <rrichter@amd.com>
Link: https://lore.kernel.org/r/166993045072.1882361.13944923741276843683.stgit@dwillia2-xfh.jf.intel.com
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The three objects 'struct cxl_nvdimm_bridge', 'struct cxl_nvdimm', and
'struct cxl_pmem_region' manage CXL persistent memory resources. The
bridge represents base platform resources, the nvdimm represents one or
more endpoints, and the region is a collection of nvdimms that
contribute to an assembled address range.
Their relationship is such that a region is torn down if any component
endpoints are removed. All regions and endpoints are torn down if the
foundational bridge device goes down.
A workqueue was deployed to manage these interdependencies, but it is
difficult to reason about, and fragile. A recent attempt to take the CXL
root device lock in the cxl_mem driver was reported by lockdep as
colliding with the flush_work() in the cxl_pmem flows.
Instead of the workqueue, arrange for all pmem/nvdimm devices to be torn
down immediately and hierarchically. A similar change is made to both
the 'cxl_nvdimm' and 'cxl_pmem_region' objects. For bisect-ability both
changes are made in the same patch which unfortunately makes the patch
bigger than desired.
Arrange for cxl_memdev and cxl_region to register a cxl_nvdimm and
cxl_pmem_region as a devres release action of the bridge device.
Additionally, include a devres release action of the cxl_memdev or
cxl_region device that triggers the bridge's release action if an endpoint
exits before the bridge. I.e. this allows either unplugging the bridge,
or unplugging and endpoint to result in the same cleanup actions.
To keep the patch smaller the cleanup of the now defunct workqueue
infrastructure is saved for a follow-on patch.
Tested-by: Robert Richter <rrichter@amd.com>
Link: https://lore.kernel.org/r/166993041773.1882361.16444301376147207609.stgit@dwillia2-xfh.jf.intel.com
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The port scanning algorithm in devm_cxl_enumerate_ports() walks up the
topology and adds cxl_port objects starting from the root down to the
endpoint. When those ports are initially created they know all their
dports, but they do not know the downstream cxl_port instance that
represents the next descendant in the topology. Rework create_endpoint()
into devm_cxl_add_endpoint() that enumerates the downstream cxl_port
topology into each port's 'struct cxl_ep' record for each endpoint it
that the port is an ancestor.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/20220624041950.559155-7-dan.j.williams@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
At the time that cxl_port instances are being created, cache the dport
from the parent port that points to this new child port. This will be
useful for region provisioning when walking the tree to calculate
decoder targets, and saves rewalking the dport list after the fact to
build this information.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/20220624041950.559155-1-dan.j.williams@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Dump the device-physical-address map for a CXL expander in /proc/iomem
style format. E.g.:
cat /sys/kernel/debug/cxl/mem1/dpamem
00000000-0fffffff : ram
10000000-1fffffff : pmem
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165603885318.551046.8308248564880066726.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The device is created, and then there is a check if a driver succesfully
bound to it. In event of failing the bind (e.g. failure in cxl_port_probe())
the device is left registered. When a bus rescan later occurs, fresh
devices are created leading to a multiple device representing the same
underlying hardware. Bad things may follow and at very least we have far too many
devices.
Fix by ensuring autoremove is registered if the device create succeeds,
but doesn't depend on sucessful binding to a driver.
Bug was observed as side effect of incorrect ownership in
[PATCH v9 6/9] cxl/port: Read CDAT table
but will result from any failure to in cxl_port_probe().
Fixes: 8dd2bc0f8e ("cxl/mem: Add the cxl_mem driver")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/20220609134519.11668-1-Jonathan.Cameron@huawei.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The responsibility for establishing HDM Decoder Capability based
operation is more closely tied to port enabling than memdev enabling
which is concerned with port enumeration. This later enables reusing
@cxlhdm for probing / controlling "global enable" for the HDM Decoder
Capability. For now, just do the nominal move.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291691167.1426646.7936109077255288258.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Now that nothing external to cxl_hdm_decode_init() considers
'struct cxl_endpoint_dvec_info' move it internal to
cxl_hdm_decode_init().
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291690612.1426646.7866084245521113414.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for changing how the driver handles 'mem_enable' in the CXL
DVSEC control register. Merge the contents of cxl_hdm_decode_init() into
cxl_dvsec_ranges() and rename the combined function cxl_hdm_decode_init().
The possible cleanups and fixes that result from this merge are saved for a
follow-on change.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/165291690027.1426646.10249756632415633752.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
When a device does not have mem_enable set then the current range
settings are moot. Skip the enumeration and cause cxl_hdm_decode_init()
to proceed directly to enable the HDM Decoder Capability.
Fixes: 560f785590 ("cxl/pci: Retrieve CXL DVSEC memory info")
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291689442.1426646.18012291761753694336.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for fixing the setting of the 'mem_enabled' bit in CXL
DVSEC Control register, move all CXL DVSEC range enumeration into the
same source file.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291688886.1426646.15046138604010482084.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Allow cxl_await_media_ready() to be mocked for testing purposes rather
than carrying the maintenance burden of an indirect function call in the
mainline driver.
With the move cxl_await_media_ready() can no longer reuse the mailbox
timeout override, so add a media_ready_timeout module parameter to the
core to backfill.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291688340.1426646.4755627801983775011.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for validating DVSEC ranges against the platform declared
CXL memory ranges (ACPI CFMWS) move port enumeration before the
endpoint's decoder validation. Ultimately this logic will move to the
port driver, but create a bisect point before that larger move.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291687749.1426646.18091538443879226995.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The addition of cxl_mem_active() broke error exit scenarios for
cxl_mem_probe(). Return early rather than proceed with disabling
suspend, and update the label name since it is no longer a terminal
"out" label that exits the function.
Fixes: 9ea4dcf498 ("PM: CXL: Disable suspend")
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291687176.1426646.15449254938752532784.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Now that wait_for_media() does nothing supplemental to
wait_for_media_ready() just promote wait_for_media_ready() to a common
helper and drop wait_for_media().
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291686046.1426646.4390664747934592185.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Media ready is asserted by the device independent of whether mem_enabled
was ever set. Drop this check to allow for dropping wait_for_media() in
favor of ->wait_media_ready().
Fixes: 8dd2bc0f8e ("cxl/mem: Add the cxl_mem driver")
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/165291685501.1426646.10372821863672431074.stgit@dwillia2-xfh
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Now that all CXL subsystem locking is validated with custom lock
classes, there is no need for the custom usage of the lockdep_mutex.
Cc: Alison Schofield <alison.schofield@intel.com>
Cc: Vishal Verma <vishal.l.verma@intel.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Ben Widawsky <ben.widawsky@intel.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/165055520383.3745911.53447786039115271.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The CXL specification claims S3 support at a hardware level, but at a
system software level there are some missing pieces. Section 9.4 (CXL
2.0) rightly claims that "CXL mem adapters may need aux power to retain
memory context across S3", but there is no enumeration mechanism for the
OS to determine if a given adapter has that support. Moreover the save
state and resume image for the system may inadvertantly end up in a CXL
device that needs to be restored before the save state is recoverable.
I.e. a circular dependency that is not resolvable without a third party
save-area.
Arrange for the cxl_mem driver to fail S3 attempts. This still nominaly
allows for suspend, but requires unbinding all CXL memory devices before
the suspend to ensure the typical DRAM flow is taken. The cxl_mem unbind
flow is intended to also tear down all CXL memory regions associated
with a given cxl_memdev.
It is reasonable to assume that any device participating in a System RAM
range published in the EFI memory map is covered by aux power and
save-area outside the device itself. So this restriction can be
minimized in the future once pre-existing region enumeration support
arrives, and perhaps a spec update to clarify if the EFI memory map is
sufficent for determining the range of devices managed by
platform-firmware for S3 support.
Per Rafael, if the CXL configuration prevents suspend then it should
fail early before tasks are frozen, and mem_sleep should stop showing
'mem' as an option [1]. Effectively CXL augments the platform suspend
->valid() op since, for example, the ACPI ops are not aware of the CXL /
PCI dependencies. Given the split role of platform firmware vs OS
provisioned CXL memory it is up to the cxl_mem driver to determine if
the CXL configuration has elements that platform firmware may not be
prepared to restore.
Link: https://lore.kernel.org/r/CAJZ5v0hGVN_=3iU8OLpHY3Ak35T5+JcBM-qs8SbojKrpd0VXsA@mail.gmail.com [1]
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Len Brown <len.brown@intel.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/165066828317.3907920.5690432272182042556.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
cxl_mem_probe() already emits a log message when HDM operation can not
be established. Delete the similar one in cxl_hdm_decode_init().
What is less obvious is why global_ctrl being enabled makes positive
values of info->ranges irrelevant, and the Linux behavior with respect
to the spec recommendation to mirror CXL Range registers with HDM
Decoder Base + Size registers.
Cc: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Link: https://lore.kernel.org/r/164944616743.454665.7055846627973202403.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
cxl_dvsec_decode_init() is tasked with checking whether legacy DVSEC
range based decode is in effect, or whether HDM can be enabled / already
is enabled. As such it either succeeds or fails and that result is the
return value. The @do_hdm_init variable is misleading in the case where
HDM operation is already found to be active, so just call it @retval.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Link: https://lore.kernel.org/r/164730736435.3806189.2537160791687837469.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for the cxl_pci driver to continue operation after
cxl_dvsec_range() failure, update cxl_mem to check for negative error
codes in info->ranges. Treat that condition as fatal regardless of the
state of the HDM configuration since cxl_mem needs positive confirmation
that legacy ranges were not established by platform firmware or another
agent.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com.
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Link: https://lore.kernel.org/r/164730735324.3806189.4167509857771192422.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
When the driver finds legacy DVSEC ranges active on a CXL Memory
Expander it indicates that platform firmware is not aware of, or is
deliberately disabling common CXL 2.0 operation. In this case Linux
generally has no choice, but to leave the device alone.
The driver attempts to validate that the DVSEC range is in the EFI
memory map. Remove that logic since there is no requirement that the
BIOS publish DVSEC ranges in the EFI Memory Map.
In the future the driver will want to permanently reserve this capacity
out of the available CFMWS capacity and hide it from
request_free_mem_region(), but it serves no purpose to warn about the
range not appearing in the EFI Memory Map.
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Link: https://lore.kernel.org/r/164730734246.3806189.13995924771963139898.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
At this point the subsystem can enumerate all CXL ports (CXL.mem decode
resources in upstream switch ports and host bridges) in a system. The
last mile is connecting those ports to endpoints.
The cxl_mem driver connects an endpoint device to the platform CXL.mem
protoctol decode-topology. At ->probe() time it walks its
device-topology-ancestry and adds a CXL Port object at every Upstream
Port hop until it gets to CXL root. The CXL root object is only present
after a platform firmware driver registers platform CXL resources. For
ACPI based platform this is managed by the ACPI0017 device and the
cxl_acpi driver.
The ports are registered such that disabling a given port automatically
unregisters all descendant ports, and the chain can only be registered
after the root is established.
Given ACPI device scanning may run asynchronously compared to PCI device
scanning the root driver is tasked with rescanning the bus after the
root successfully probes.
Conversely if any ports in a chain between the root and an endpoint
becomes disconnected it subsequently triggers the endpoint to
unregister. Given lock depenedencies the endpoint unregistration happens
in a workqueue asynchronously. If userspace cares about synchronizing
delayed work after port events the /sys/bus/cxl/flush attribute is
available for that purpose.
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
[djbw: clarify changelog, rework hotplug support]
Link: https://lore.kernel.org/r/164398782997.903003.9725273241627693186.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
As the driver has undergone development, it's become clear that the
majority [entirety?] of the current functionality in mem.c is actually a
layer encapsulating functionality exposed through PCI based
interactions. This layer can be used either in isolation or to provide
functionality for higher level functionality.
CXL capabilities exist in a parallel domain to PCIe. CXL devices are
enumerable and controllable via "legacy" PCIe mechanisms; however, their
CXL capabilities are a superset of PCIe. For example, a CXL device may
be connected to a non-CXL capable PCIe root port, and therefore will not
be able to participate in CXL.mem or CXL.cache operations, but can still
be accessed through PCIe mechanisms for CXL.io operations.
To properly represent the PCI nature of this driver, and in preparation for
introducing a new driver for the CXL.mem / HDM decoder (Host-managed Device
Memory) capabilities of a CXL memory expander, rename mem.c to pci.c so that
mem.c is available for this new driver.
The result of the change is that there is a clear layering distinction
in the driver, and a systems administrator may load only the cxl_pci
module and gain access to such operations as, firmware update, offline
provisioning of devices, and error collection. In addition to freeing up
the file name for another purpose, there are two primary reasons this is
useful,
1. Acting upon devices which don't have full CXL capabilities. This
may happen for instance if the CXL device is connected in a CXL
unaware part of the platform topology.
2. Userspace-first provisioning for devices without kernel driver
interference. This may be useful when provisioning a new device
in a specific manner that might otherwise be blocked or prevented
by the real CXL mem driver.
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Ben Widawsky <ben.widawsky@intel.com>
Link: https://lore.kernel.org/r/20210526174413.802913-1-ben.widawsky@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
While CXL Memory Device endpoints locate the CXL MMIO registers in a PCI
BAR, CXL root bridges have their MMIO base address described by platform
firmware. Refactor the existing register lookup into a generic facility
for endpoints and bridges to share.
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/162096972534.1865304.3218686216153688039.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
CXL MMIO register blocks are organized by device type and capabilities.
There are Component registers, Device registers (yes, an ambiguous
name), and Memory Device registers (a specific extension of Device
registers).
It is possible for a given device instance (endpoint or port) to
implement register sets from multiple of the above categories.
The driver code that enumerates and maps the registers is type specific
so it is useful to have a dedicated type and helpers for each block
type.
At the same time, once the registers are mapped the origin type does not
matter. It is overly pedantic to reference the register block type in
code that is using the registers.
In preparation for the endpoint driver to incorporate Component registers
into its MMIO operations reorganize the registers to allow typed
enumeration + mapping, but anonymous usage. With the end state of
'struct cxl_regs' to be:
struct cxl_regs {
union {
struct {
CXL_DEVICE_REGS();
};
struct cxl_device_regs device_regs;
};
union {
struct {
CXL_COMPONENT_REGS();
};
struct cxl_component_regs component_regs;
};
};
With this arrangement the driver can share component init code with
ports, but when using the registers it can directly reference the
component register block type by name without the 'component_regs'
prefix.
So, map + enumerate can be shared across drivers of different CXL
classes e.g.:
void cxl_setup_device_regs(struct device *dev, void __iomem *base,
struct cxl_device_regs *regs);
void cxl_setup_component_regs(struct device *dev, void __iomem *base,
struct cxl_component_regs *regs);
...while inline usage in the driver need not indicate where the
registers came from:
readl(cxlm->regs.mbox + MBOX_OFFSET);
readl(cxlm->regs.hdm + HDM_OFFSET);
...instead of:
readl(cxlm->regs.device_regs.mbox + MBOX_OFFSET);
readl(cxlm->regs.component_regs.hdm + HDM_OFFSET);
This complexity of the definition in .h yields improvement in code
readability in .c while maintaining type-safety for organization of
setup code. It prepares the implementation to maintain organization in
the face of CXL devices that compose register interfaces consisting of
multiple types.
Given that this new container is named 'regs' rename the common register
base pointer @base, and fixup the kernel-doc for the missing @cxlmd
description.
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/162096971451.1865304.13540251513463515153.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for sharing cxl.h with other generic CXL consumers,
move / consolidate some of the memory device specifics to mem.h.
The motivation for moving out of cxl.h is to maintain least privilege
access to memory-device details since cxl.h is used in multiple files.
The motivation for moving definitions into a new mem.h header is for
code readability and organization. I.e. minimize implementation details
when reading data structures and other definitions.
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/162096970932.1865304.14510894426562947262.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The CXL Identify Memory Device output payload emits capacity in 256MB
units. The driver is treating the capacity field as bytes. This was
missed because QEMU reports bytes when it should report bytes / 256MB.
Fixes: 8adaf747c9 ("cxl/mem: Find device capabilities")
Reviewed-by: Vishal Verma <vishal.l.verma@intel.com>
Cc: Ben Widawsky <ben.widawsky@intel.com>
Link: https://lore.kernel.org/r/161862021044.3259705.7008520073059739760.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The "Register Offset Low" register of a "DVSEC Register Locator"
contains the 64K aligned offset for the registers along with the BAR
indicator and an id. The implementation was treating the "Register Block
Offset Low" field a value rather than as a pre-aligned component of the
64-bit offset. So, just mask, don't mask and shift (FIELD_GET).
The user visible result of this bug is that the driver fails to bind to
the device after none of the required blocks are found.
This was missed earlier because the primary development done in the QEMU
environment only uses 0 offsets, i.e. 0 shifted is still 0.
Fixes: 8adaf747c9 ("cxl/mem: Find device capabilities")
Reported-by: Vishal Verma <vishal.l.verma@intel.com>
Signed-off-by: Ben Widawsky <ben.widawsky@intel.com>
Link: https://lore.kernel.org/r/20210415232610.603273-1-ben.widawsky@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Typically the mem_commands[] array is in sync with 'enum { CXL_CMDS }'.
Current code works well.
However, the array size of mem_commands[] may not strictly be the same
as CXL_MEM_COMMAND_ID_MAX. E.g. if a new CXL_CMD() is added that is
guarded by #ifdefs, the array could be shorter. This could lead then
further to an out-of-bounds array access in cxl_validate_cmd_from_user().
Fix this by forcing the array size to CXL_MEM_COMMAND_ID_MAX. This
also adds range checks for array items in mem_commands[] at compile
time.
Signed-off-by: Robert Richter <rrichter@amd.com>
Link: https://lore.kernel.org/r/20210324141635.22335-1-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
There is no power management of cxl virtual devices, disable
device-power-management and runtime-power-management to prevent
userspace from growing expectations of those attributes appearing. They
can be added back in the future if needed.
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Link: https://lore.kernel.org/r/161728761025.2474381.808344500111924819.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
While device_add() will happen to catch dev_set_name() failures it is a
broken pattern to follow given that the core may try to fall back to a
different name.
Add explicit checking for dev_set_name() failures to be cleaned up by
put_device(). Skip cdev_device_add() and proceed directly to
put_device() if the name set fails.
This type of bug is easier to see if 'alloc' is split from 'add'
operations that require put_device() on failure. So cxl_memdev_alloc()
is split out as a result.
Fixes: b39cb1052a ("cxl/mem: Register CXL memX devices")
Reported-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/161728760514.2474381.1163928273337158134.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The percpu_ref to gate whether cxl_memdev_ioctl() is free to use the
driver context (@cxlm) to issue I/O is overkill, implemented incorrectly
(missing a device reference before accessing the percpu_ref), and the
complexities of shutting down a percpu_ref contributed to a bug in the
error unwind in cxl_mem_add_memdev() (missing put_device() to be fixed
separately).
Use an rwsem to explicitly synchronize the usage of cxlmd->cxlm, and add
the missing reference counting for cxlmd in cxl_memdev_open() and
cxl_memdev_release_file().
Fixes: b39cb1052a ("cxl/mem: Register CXL memX devices")
Reported-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/161728759948.2474381.17481500816783671817.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
While none the CXL sysfs attributes are threatening to overrun a
PAGE_SIZE of output, it is good form to use the recommended helpers.
Fixes: b39cb1052a ("cxl/mem: Register CXL memX devices")
Reported-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/161728759424.2474381.11231441014951343463.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
When submitting a command for userspace, input and output payload bounce
buffers are allocated. For a given command, both input and output
buffers may exist and so when allocation of the input buffer fails, the
output buffer must be freed too.
As far as I can tell, userspace can't easily exploit the leak to OOM a
machine unless the machine was already near OOM state.
Fixes: 583fa5e71c ("cxl/mem: Add basic IOCTL interface")
Reported-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Link: https://lore.kernel.org/r/20210221035846.680145-1-ben.widawsky@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The copy_to_user() function returns the number of bytes remaining to be
copied, but we want to return -EFAULT if the copy doesn't complete.
Fixes: b754ffbbc0ee ("cxl/mem: Add basic IOCTL interface")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/YC+K3kgzqm20zCWY@mwanda
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Add initial set of formal commands beyond basic identify and command
enumeration.
Signed-off-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> (v2)
Link: https://lore.kernel.org/r/20210217040958.1354670-8-ben.widawsky@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
